The binding of a fully flexible peptide, loop, or terminus of a protein to a relatively structured site on the surface of its molecular partner is a prevalent type of a transient macromolecular interaction that mediates the majority of recognition and signaling cascades. While spanning a wide range of affinities, such interactions uniformly involve a structured and an unstructured partner that may exist as independent entities in the absence of one another and only associate for a limited period of time resulting in conformational changes, recruitment or dissociation of other proteins to/from the complex, and activation or inhibition of the downstream pathways. Synthetic peptide modulators of protein-peptide interactions are frequently promising therapeutic candidates in cancer, inflammation and endocrine disorders. Despite the critical importance of transient protein-peptide interactions for biomedical research and therapeutic discovery, only a small fraction of these complexes are amenable to experimental structure determination. Therefore, only accurate peptide docking may lead to a breakthrough in structural understanding of protein-peptide interactions. Yet for peptides longer than 6-8 amino acids, the overwhelming size and complexity of the conformational search space, the inevitable inaccuracies in the binding site representation due to induced fit, inefficient or insufficiently thorough sampling, and accumulation of scoring function errors prohibit the accurate determination of peptide binding poses and interactions by computational methods. The present proposal intends to dramatically expand the range of peptide and protein sizes for which accurate complex geometry prediction can be achieved by global conformational optimization. This advance will be made by pursuing two Specific Aims: (Aim 1) Development and optimization of a reliable peptide cross-docking procedure using chemical field-enhanced binding site representations and improved entropy calculations in the stochastic global conformational search in internal coordinates; and (Aim 2) Extension of the developed protocol to specific biological projects with experimental validation of the predicted geometries and peptide variants. The innovative strategies proposed in Aim 1 include using a new force field, the enrichment of binding site representations with chemical fields, and the optimal conformational expansion of the pocket for induced fit. The targets of Aim 2 include complexes of Class A and B GPCRs with their protein and peptide modulators and G-protein ?i interactions with GEF peptides. The attainment of the aims of this proposal will lead to breakthrough advances in flexible macromolecular docking. It will result in valuable software tools, protocols and shared resources for the biological community. It will also lead to discovery of new peptide modulators of important therapeutic, immune, and diagnostic targets.